INDEX

Explanations

text related to corrections or accuracy

oai_token-act-pair · gpt-3.5-turbo

references to the act of correcting or making amendments

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

atos

-0.77

aden

-0.73

 carrot

-0.73

thin

-0.72

nam

-0.69

groups

-0.65

NetMessage

-0.65

thood

-0.63

Connector

-0.61

warm

-0.61

POSITIVE LOGITS

 erroneous

0.98

 misinformation

0.97

 spelling

0.95

ions

0.90

 inaccur

0.87

orate

0.85

 errors

0.83

 misconceptions

0.81

itude

0.79

 terminology

0.78

Activations Density 0.048%

text related to corrections or accuracy

references to the act of correcting or making amendments

No Comments

No Known Activations