INDEX

Explanations

phrases related to methods or steps

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate methods or approaches to achieve a specific outcome

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

usters

-0.86

asts

-0.74

qus

-0.70

ighed

-0.69

noxious

-0.69

vertisement

-0.68

enture

-0.68

egu

-0.66

Interstitial

-0.66

irie

-0.66

POSITIVE LOGITS

 forward

1.09

to

0.98

forward

0.87

 Forward

0.86

 forwards

0.78

TO

0.73

ever

0.68

out

0.67

vernment

0.64

you

0.64

Activations Density 0.055%

phrases related to methods or steps

phrases that indicate methods or approaches to achieve a specific outcome

No Comments

No Known Activations