INDEX

Explanations

statements indicating errors or the concept of making mistakes

oai_token-act-pair · gpt-3.5-turbo

mentions of mistakes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

electric

-0.79

ighth

-0.79

well

-0.78

rollers

-0.78

ingers

-0.73

accompan

-0.70

tsky

-0.70

idal

-0.68

undai

-0.68

amen

-0.68

POSITIVE LOGITS

 mistake

1.09

 unfocusedRange

1.02

 mistakes

0.97

ously

0.96

 mistaken

0.84

 Malf

0.79

 errors

0.78

 error

0.78

é»Ĵ

0.75

Maker

0.74

Activations Density 0.018%

statements indicating errors or the concept of making mistakes

mentions of mistakes

No Comments

No Known Activations