INDEX

Explanations

numerical values related to scores in a sports context

oai_token-act-pair · gpt-3.5-turbo

instances of the word "scored," typically in the context of sports or achievements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

agan

-0.85

lly

-0.72

UAL

-0.67

ledge

-0.62

ually

-0.61

ned

-0.61

 compulsion

-0.61

ships

-0.59

por

-0.59

how

-0.59

POSITIVE LOGITS

 scored

1.11

 scores

0.99

 score

0.98

 scoring

0.86

iage

0.84

ificant

0.81

otle

0.80

card

0.76

 Scores

0.76

umbles

0.76

Activations Density 0.010%

numerical values related to scores in a sports context

instances of the word "scored," typically in the context of sports or achievements

No Comments

No Known Activations