INDEX

Explanations

words related to conflict, specifically ambush situations

oai_token-act-pair · gpt-3.5-turbo

terms related to ambiguity and unclear situations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Occupations

-0.77

 Parenthood

-0.72

Material

-0.70

Fuel

-0.69

fork

-0.69

ppo

-0.68

leaf

-0.68

é¾įå¥ĳå£«

-0.67

ãĥīãĥ©ãĤ´ãĥ³

-0.67

 AMERICA

-0.67

POSITIVE LOGITS

ience

0.99

ushed

0.97

assad

0.96

ivalent

0.93

isexual

0.92

iences

0.91

adel

0.89

iov

0.89

cade

0.88

ushes

0.88

Activations Density 0.016%

words related to conflict, specifically ambush situations

terms related to ambiguity and unclear situations

No Comments

No Known Activations