INDEX

Explanations

phrases related to long-term plans or concepts

oai_token-act-pair · gpt-3.5-turbo

references to long-term policies and commitments

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

 Scrib

-0.90

 GEAR

-0.85

 Compass

-0.83

AVG

-0.82

ulhu

-0.79

Wiz

-0.77

 DRAG

-0.76

ħĭ

-0.75

Phi

-0.73

Ö¼

-0.72

POSITIVE LOGITS

distance

1.14

term

1.13

enough

1.12

sighted

1.05

haired

1.05

sized

1.05

eyed

1.03

sounding

1.02

serving

1.00

suff

1.00

Activations Density 0.043%

phrases related to long-term plans or concepts

references to long-term policies and commitments

No Comments

No Known Activations