INDEX

Explanations

references to baseball championships, particularly the World Series

oai_token-act-pair · gpt-3.5-turbo

references to various sports championships

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

anyahu

-0.82

zsche

-0.74

oidal

-0.73

atem

-0.73

tical

-0.72

 scanner

-0.71

asonic

-0.70

body

-0.70

organic

-0.70

axis

-0.70

POSITIVE LOGITS

 Finals

1.03

 Playoff

0.99

 Playoffs

0.92

 Trophy

0.85

 Baseball

0.84

Era

0.83

 Quarter

0.82

å£«

0.81

 playoffs

0.81

 victories

0.78

Activations Density 0.042%

references to baseball championships, particularly the World Series

references to various sports championships

No Comments

No Known Activations