INDEX

Explanations

references to walking or movement

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the word "walk" in various forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

encies

-0.78

iled

-0.68

afort

-0.66

nan

-0.65

iling

-0.65

nces

-0.65

igne

-0.63

cffff

-0.62

orporated

-0.62

iller

-0.61

POSITIVE LOGITS

through

1.01

about

0.91

 upright

0.90

 away

0.89

ways

0.86

 uphill

0.80

 onstage

0.79

 bare

0.78

way

0.76

bow

0.75

Activations Density 0.032%

references to walking or movement

repeated instances of the word "walk" in various forms

No Comments

No Known Activations