INDEX

Explanations

names of players and clubs associated with the football/soccer team Real Madrid

oai_token-act-pair · gpt-3.5-turbo

references to Real Madrid and its star player Cristiano Ronaldo

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ohl

-0.77

awk

-0.75

REE

-0.74

RON

-0.73

pee

-0.72

tail

-0.72

raid

-0.72

 Dickinson

-0.70

AIN

-0.69

cutting

-0.68

POSITIVE LOGITS

 Ronaldo

1.10

 Madrid

0.91

atu

0.73

 Messi

0.71

ventus

0.71

 Liga

0.70

igne

0.68

 athlet

0.68

ombies

0.68

Caf

0.67

Activations Density 0.067%

names of players and clubs associated with the football/soccer team Real Madrid

references to Real Madrid and its star player Cristiano Ronaldo

No Comments

No Known Activations