INDEX

Explanations

instructions or steps in technical content, specifically related to adding or loading

oai_token-act-pair · gpt-3.5-turbo

technical terms related to computing or programming

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ember

-0.82

issions

-0.78

uru

-0.73

ĸļ

-0.73

dash

-0.71

ittance

-0.69

mons

-0.68

uble

-0.66

ãĥģ

-0.65

unic

-0.64

POSITIVE LOGITS

 disg

0.64

 overseen

0.64

herty

0.63

orts

0.59

 biased

0.59

 rife

0.57

 Prol

0.57

gol

0.57

 poisoned

0.56

 skewed

0.56

Activations Density 0.000%

instructions or steps in technical content, specifically related to adding or loading

technical terms related to computing or programming

No Comments

No Known Activations