INDEX

Explanations

directional words related to geographic locations

oai_token-act-pair · gpt-3.5-turbo

directional words related to movement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 nutshell

-0.68

iqueness

-0.63

 peppers

-0.62

ollar

-0.62

urrection

-0.61

ggles

-0.61

ettle

-0.61

ulous

-0.60

 popup

-0.60

uracy

-0.60

POSITIVE LOGITS

 toward

1.07

 towards

1.02

wards

0.98

WARD

0.88

ward

0.85

 unnoticed

0.83

 stairs

0.79

 into

0.78

travel

0.78

wind

0.76

Activations Density 0.146%

directional words related to geographic locations

directional words related to movement

No Comments

No Known Activations