INDEX

Explanations

proper nouns or titles related to sports

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual, possibly an athlete or public figure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Amen

-0.77

endif

-0.74

peak

-0.71

 Capitalism

-0.68

edia

-0.66

Faw

-0.64

ombat

-0.64

 Proof

-0.64

 Online

-0.64

 Simpl

-0.64

POSITIVE LOGITS

 teammates

0.86

eded

0.83

 rookie

0.82

inous

0.81

 teammate

0.80

 athletic

0.79

pherd

0.78

 athleticism

0.77

 coaches

0.77

 undrafted

0.77

Activations Density 0.216%

proper nouns or titles related to sports

references to a specific individual, possibly an athlete or public figure

No Comments

No Known Activations