INDEX

Explanations

statements of certainty based on evidence

oai_token-act-pair · gpt-3.5-turbo

phrases indicating certainty or conclusions

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

usk

-0.72

ļéĨĴ

-0.71

dozen

-0.69

urai

-0.65

 pione

-0.65

ocks

-0.63

yna

-0.63

incible

-0.62

ãĤ´ãĥ³

-0.62

Review

-0.62

POSITIVE LOGITS

 there

1.10

 something

1.06

 whoever

0.99

 they

0.97

it

0.90

 somebody

0.87

 someone

0.87

 this

0.84

 everything

0.82

 THEY

0.82

Activations Density 0.283%

statements of certainty based on evidence

phrases indicating certainty or conclusions

No Comments

No Known Activations