INDEX

Explanations

specific noun phrases related to legal or political entities

oai_token-act-pair · gpt-3.5-turbo

references to entities or groups related to legal or political designations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

nikov

-0.77

raft

-0.76

ictional

-0.75

ancial

-0.69

Stage

-0.69

entanyl

-0.68

abbling

-0.67

arate

-0.67

raltar

-0.66

estic

-0.66

POSITIVE LOGITS

 designers

0.86

 defenders

0.84

 planners

0.82

ĳå£«

0.82

 deems

0.81

 observers

0.79

 humans

0.79

 admins

0.78

 pundits

0.77

 readers

0.77

Activations Density 0.211%

specific noun phrases related to legal or political entities

references to entities or groups related to legal or political designations

No Comments

No Known Activations