INDEX

Explanations

instances of misrepresentation or falsification

oai_token-act-pair · gpt-3.5-turbo

terms related to misrepresentation and falsification

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

obbies

-0.70

 flush

-0.68

raid

-0.66

STON

-0.66

qua

-0.65

thood

-0.65

MRI

-0.65

NetMessage

-0.65

aternity

-0.64

 avid

-0.63

POSITIVE LOGITS

 misrepresent

1.26

ations

1.17

ation

1.05

ed

0.92

edly

0.84

ifying

0.80

ified

0.78

ifications

0.78

rities

0.77

ments

0.77

Activations Density 0.007%

instances of misrepresentation or falsification

terms related to misrepresentation and falsification

No Comments

No Known Activations