INDEX

Explanations

strongly worded declarations of authorial opinions or arguments

oai_token-act-pair · gpt-3.5-turbo

the word "that" indicating arguments or claims

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

backer

-0.80

stal

-0.70

Guard

-0.67

mouth

-0.65

api

-0.63

SEE

-0.60

inar

-0.60

Champ

-0.60

atro

-0.60

aq

-0.60

POSITIVE LOGITS

 there

0.75

 although

0.72

 justifies

0.69

 preserving

0.69

 abol

0.67

 someday

0.67

we

0.65

 whoever

0.65

 prevailed

0.65

"[

0.64

Activations Density 0.217%

strongly worded declarations of authorial opinions or arguments

the word "that" indicating arguments or claims

No Comments

No Known Activations