INDEX

Explanations

phrases related to walking

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "walking" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nan

-0.75

ucks

-0.73

 Aerospace

-0.72

cffff

-0.72

ional

-0.71

nces

-0.70

unal

-0.69

ilies

-0.68

iled

-0.68

illion

-0.68

POSITIVE LOGITS

own

0.80

 upright

0.76

 tours

0.75

 Walk

0.75

walking

0.74

 Tours

0.71

DragonMagazine

0.70

walker

0.69

 Guides

0.69

 away

0.69

Activations Density 0.022%

phrases related to walking

occurrences of the word "walking" in various contexts

No Comments

No Known Activations