INDEX

Explanations

sports-related terms, especially related to the NFL and NBA

oai_token-act-pair · gpt-3.5-turbo

references to professional sports leagues, particularly the NFL and NHL

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

zl

-0.77

AFB

-0.77

 Gork

-0.71

chin

-0.70

romy

-0.69

 Shepard

-0.68

Yen

-0.68

akov

-0.68

_>

-0.67

Ae

-0.66

POSITIVE LOGITS

 Playoff

0.98

 Playoffs

0.97

 leagues

0.95

 Franch

0.94

League

0.93

 offic

0.92

 rosters

0.91

 referees

0.91

 Championship

0.91

 franch

0.87

Activations Density 0.240%

sports-related terms, especially related to the NFL and NBA

references to professional sports leagues, particularly the NFL and NHL

No Comments

No Known Activations