INDEX

Explanations

mentions of the word "game"

oai_token-act-pair · gpt-3.5-turbo

references to the word "game."

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

thening

-0.81

thens

-0.77

cellaneous

-0.73

orescent

-0.73

hips

-0.72

inois

-0.71

 unfocusedRange

-0.66

 politic

-0.66

ayers

-0.65

sie

-0.65

POSITIVE LOGITS

pad

1.08

wright

1.06

 Borderlands

1.03

PLAY

0.92

cube

0.89

arium

0.87

play

0.85

 Ubisoft

0.84

 consoles

0.84

 Thief

0.80

Activations Density 0.057%

mentions of the word "game"

references to the word "game."

No Comments

No Known Activations