INDEX

Explanations

words related to actions, decisions, and behaviors, with a focus on behavior towards tasks and responsibilities

oai_token-act-pair · gpt-3.5-turbo

verbs indicating actions or decisions related to personal or psychological states

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

selves

-0.67

 Composite

-0.61

mination

-0.58

earch

-0.58

 respectively

-0.58

 unison

-0.57

 discrep

-0.57

 guiActiveUnfocused

-0.56

ogether

-0.55

Else

-0.54

POSITIVE LOGITS

 himself

1.07

 Himself

0.69

 personally

0.65

pocket

0.58

 directing

0.56

panic

0.55

 reelection

0.54

his

0.54

liest

0.54

 wife

0.53

Activations Density 1.270%

words related to actions, decisions, and behaviors, with a focus on behavior towards tasks and responsibilities

verbs indicating actions or decisions related to personal or psychological states

No Comments

No Known Activations