INDEX

Explanations

quotations or reported speech

oai_token-act-pair · gpt-3.5-turbo

expressions of dialogue or quotations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 prohib

-0.74

 piled

-0.62

 pole

-0.61

irtual

-0.60

 habitable

-0.59

 interpre

-0.59

enary

-0.59

 nearest

-0.58

 similarly

-0.58

Lau

-0.57

POSITIVE LOGITS

YES

0.88

 congratulations

0.88

hey

0.88

Oh

0.86

bye

0.84

 prest

0.81

HHHH

0.80

bye

0.80

sir

0.80

GOD

0.78

Activations Density 0.212%

quotations or reported speech

expressions of dialogue or quotations

No Comments

No Known Activations