INDEX

Explanations

foreign characters, potentially indicating specific coding or language-related patterns

oai_token-act-pair · gpt-3.5-turbo

symbols and punctuation in the context of text formatting or coding

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Bacon

-0.96

Bag

-0.94

XI

-0.94

Pes

-0.93

Eps

-0.93

 Grill

-0.92

 Union

-0.92

 Opinion

-0.92

 Meter

-0.92

 Xiao

-0.91

POSITIVE LOGITS

subject

1.56

nob

1.54

future

1.53

general

1.49

same

1.47

average

1.47

new

1.46

simple

1.46

license

1.46

super

1.46

Activations Density 0.628%

foreign characters, potentially indicating specific coding or language-related patterns

symbols and punctuation in the context of text formatting or coding

No Comments

No Known Activations