INDEX

Explanations

phrases asking for specific information or clarification

oai_token-act-pair · gpt-3.5-turbo

queries that seek clarification or explanation about specific topics

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

Runner

-0.75

mur

-0.74

bis

-0.72

gi

-0.71

oco

-0.71

adiq

-0.70

mates

-0.69

bart

-0.67

otos

-0.66

gio

-0.66

POSITIVE LOGITS

 constitutes

1.29

 transpired

1.27

 happened

1.15

 happens

1.10

 qualifies

1.04

 distinguishes

0.96

 entails

0.95

 separates

0.91

 motiv

0.88

 bothers

0.83

Activations Density 0.073%

phrases asking for specific information or clarification

queries that seek clarification or explanation about specific topics

No Comments

No Known Activations