INDEX

Explanations

terms related to an unresolved, outstanding, or undefeated state

oai_token-act-pair · gpt-3.5-turbo

terms related to unresolved or ongoing issues

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

udic

-0.92

oras

-0.89

azine

-0.83

asio

-0.77

ushima

-0.75

pty

-0.75

ueller

-0.74

jah

-0.73

opter

-0.72

olog

-0.71

POSITIVE LOGITS

NESS

0.77

INGTON

0.72

cies

0.65

ness

0.62

stretched

0.61

nesses

0.60

 Tribal

0.60

lihood

0.60

ingly

0.58

 curfew

0.57

Activations Density 0.091%

terms related to an unresolved, outstanding, or undefeated state

terms related to unresolved or ongoing issues

No Comments

No Known Activations