INDEX

Explanations

dates in the format (month) with high activations

oai_token-act-pair · gpt-3.5-turbo

parentheses and associated punctuation marks

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

 answ

-0.87

 guiActiveUn

-0.74

Ń·

-0.73

ãĤ©

-0.72

 mathemat

-0.71

hang

-0.67

rote

-0.67

 citiz

-0.67

omorphic

-0.64

 simul

-0.64

POSITIVE LOGITS

 Committees

0.74

 Expend

0.73

)].

0.72

 Conversion

0.71

>]

0.69

ioch

0.69

)))

0.68

eous

0.68

 Gamb

0.67

 Syndicate

0.67

Activations Density 0.095%

dates in the format (month) with high activations

parentheses and associated punctuation marks

No Comments

No Known Activations