INDEX

Explanations

names of sports players who have scored goals or points in various matches or games

oai_token-act-pair · gpt-3.5-turbo

instances of the word "scored" associated with various sports contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

-0.76

tan

-0.72

lly

-0.69

nown

-0.69

xus

-0.61

 suit

-0.61

 settlers

-0.61

xon

-0.60

 Layer

-0.60

hue

-0.60

POSITIVE LOGITS

keeper

0.94

 scored

0.90

card

0.86

 scoring

0.82

 touchdowns

0.80

heet

0.77

scoring

0.77

 Rica

0.76

keeping

0.74

keepers

0.74

Activations Density 0.013%

names of sports players who have scored goals or points in various matches or games

instances of the word "scored" associated with various sports contexts

No Comments

No Known Activations