INDEX

Explanations

mentions of actions related to music or sports events and releases

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

verbs related to creation or production

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arine

-0.68

CLOSE

-0.68

assis

-0.66

lex

-0.65

Asked

-0.63

hatt

-0.62

 Niet

-0.62

Solution

-0.61

onian

-0.61

hood

-0.61

POSITIVE LOGITS

 landfall

1.29

 headlines

1.19

 waves

1.12

 strides

1.06

 history

1.00

 appearances

0.98

hift

0.93

 cameo

0.92

 pilgr

0.84

 contact

0.82

Activations Density 0.104%

mentions of actions related to music or sports events and releases

verbs related to creation or production

No Comments

No Known Activations