INDEX

Explanations

mentions of formidable or significant entities, often with a negative connotation

oai_token-act-pair · gpt-3.5-turbo

terms related to mythical or monstrous entities

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

cius

-0.74

angan

-0.70

onica

-0.70

arters

-0.69

 eligibility

-0.66

activated

-0.65

lace

-0.63

Closure

-0.63

uren

-0.62

gemony

-0.62

POSITIVE LOGITS

 beasts

0.98

 beast

0.97

osaurs

0.87

gery

0.82

stros

0.78

ngth

0.78

bru

0.74

 Beasts

0.74

 sized

0.73

warts

0.72

Activations Density 0.104%

mentions of formidable or significant entities, often with a negative connotation

terms related to mythical or monstrous entities

No Comments

No Known Activations