INDEX

Explanations

text related to identifying, acknowledging, and resolving issues or problems for fixing

oai_token-act-pair · gpt-3.5-turbo

references to fixing or repairing issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Champ

-0.71

 sidx

-0.70

ographic

-0.70

 GOODMAN

-0.65

retion

-0.64

ĸļ

-0.64

clinton

-0.64

ographies

-0.63

aez

-0.60

é¾įå¥ĳå£«

-0.59

POSITIVE LOGITS

ated

0.93

 bugs

0.87

 broken

0.87

ate

0.86

 problems

0.86

 faults

0.84

num

0.81

 flaws

0.81

 mistakes

0.79

tle

0.79

Activations Density 0.060%

text related to identifying, acknowledging, and resolving issues or problems for fixing

references to fixing or repairing issues

No Comments

No Known Activations