INDEX

Explanations

research-related phrases and findings in scientific studies

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the presentation of research findings or claims

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

iciary

-0.87

apult

-0.83

Justice

-0.83

redit

-0.83

ocaust

-0.81

DragonMagazine

-0.81

wow

-0.80

Tank

-0.79

Draft

-0.79

adr

-0.79

POSITIVE LOGITS

 exposure

1.01

 altering

0.98

 prolonged

0.97

 combining

0.96

 ingestion

0.96

 modifying

0.95

 activating

0.94

 inhibition

0.94

 correlations

0.93

 lowering

0.93

Activations Density 0.190%

research-related phrases and findings in scientific studies

phrases indicating the presentation of research findings or claims

No Comments

No Known Activations