INDEX

Explanations

actions involving checking, examining, or evaluating something

oai_token-act-pair · gpt-3.5-turbo

phrases related to checking or observing situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

versive

-0.72

ÃŃa

-0.70

inqu

-0.66

theless

-0.65

onto

-0.63

prime

-0.62

 whistle

-0.61

ãĥĨãĤ£

-0.61

onz

-0.61

ieri

-0.60

POSITIVE LOGITS

how

1.17

 whether

1.02

 firsthand

0.94

 what

0.94

if

0.89

 whats

0.87

how

0.83

whether

0.75

what

0.74

why

0.73

Activations Density 0.074%

actions involving checking, examining, or evaluating something

phrases related to checking or observing situations

No Comments

No Known Activations