INDEX

Explanations

words related to problems, issues, and challenges tackled in various settings

oai_token-act-pair · gpt-3.5-turbo

phrases indicating problems or issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

robe

-0.80

chel

-0.76

rug

-0.75

throp

-0.75

onga

-0.74

oad

-0.73

amins

-0.71

 salute

-0.71

aunder

-0.71

RAY

-0.69

POSITIVE LOGITS

 plag

0.96

 unresolved

0.91

 arising

0.88

 worsened

0.82

 arise

0.81

 escalated

0.77

 exacerbated

0.75

 Watergate

0.75

 aggravated

0.73

 Fowler

0.73

Activations Density 0.368%

words related to problems, issues, and challenges tackled in various settings

phrases indicating problems or issues

No Comments

No Known Activations