INDEX

Explanations

words with a prefix or suffix indicating an action or change

oai_token-act-pair · gpt-3.5-turbo

occurrences of prefixes or beginnings of words that indicate actions or feelings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 consequential

-0.80

 guiActiveUnfocused

-0.76

 generic

-0.69

 secondary

-0.67

 differential

-0.67

 attribution

-0.66

 hostage

-0.65

sonian

-0.65

 inclusive

-0.65

 mandatory

-0.64

POSITIVE LOGITS

wrote

1.23

izes

1.09

ook

1.07

itates

1.07

uld

1.05

igrated

1.01

ifies

0.99

ently

0.97

cedes

0.97

saw

0.96

Activations Density 0.242%

words with a prefix or suffix indicating an action or change

occurrences of prefixes or beginnings of words that indicate actions or feelings

No Comments

No Known Activations