INDEX

Explanations

References to a specific entity called "endorf" or variations thereof

oai_token-act-pair · gpt-3.5-turbo

mentions of endorsements or individuals associated with endorsements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

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

roman

-0.62

 Primordial

-0.58

zees

-0.57

scroll

-0.56

 Period

-0.56

 Overt

-0.55

obic

-0.55

HB

-0.55

 separatists

-0.55

 unsub

-0.54

POSITIVE LOGITS

endor

1.54

ffer

0.87

abilia

0.87

lance

0.81

iture

0.79

cade

0.76

ãĥ¼ãĥĨãĤ£

0.76

enance

0.76

lette

0.73

ption

0.71

Activations Density 0.007%

References to a specific entity called "endorf" or variations thereof

mentions of endorsements or individuals associated with endorsements

No Comments

No Known Activations