INDEX

Explanations

sports-related terms and specific names of athletes or teams

oai_token-act-pair · gpt-3.5-turbo

instances of sports-related grades and evaluations

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

xual

-0.86

 explicitly

-0.74

 ultr

-0.73

 extermination

-0.73

 moder

-0.71

 popul

-0.70

 anarchism

-0.70

 arrang

-0.69

rosis

-0.68

 administr

-0.67

POSITIVE LOGITS

NBA

1.00

Game

0.98

GAME

0.97

Coach

0.96

Notable

0.95

ESPN

0.91

Players

0.91

Team

0.88

Pros

0.88

football

0.86

Activations Density 0.260%

sports-related terms and specific names of athletes or teams

instances of sports-related grades and evaluations

No Comments

No Known Activations