INDEX

Explanations

sports-related activities and events

oai_token-act-pair · gpt-3.5-turbo

contrasting phrases and qualifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

emin

-0.79

reprene

-0.77

yll

-0.74

riet

-0.73

sth

-0.71

uscript

-0.70

alogue

-0.68

TeX

-0.68

itarian

-0.67

isure

-0.66

POSITIVE LOGITS

 coached

0.95

 albeit

0.93

 Sammy

0.92

 Bengals

0.88

 rookie

0.86

 teammate

0.83

 Bobby

0.81

Team

0.81

 namely

0.79

 coach

0.78

Activations Density 0.526%

sports-related activities and events

contrasting phrases and qualifications

No Comments

No Known Activations