INDEX

Explanations

words related to actions or activities, especially those related to publication, premieres, releases, or revelations

oai_token-act-pair · gpt-3.5-turbo

past tense verbs and phrases indicating actions or events

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

-0.74

lear

-0.66

carry

-0.65

SIGN

-0.65

obin

-0.65

lie

-0.64

ilater

-0.62

essing

-0.61

 compliments

-0.58

Jr

-0.57

POSITIVE LOGITS

utenberg

0.77

pload

0.70

 premier

0.68

mite

0.66

SPI

0.65

PSP

0.64

metics

0.64

psc

0.63

dule

0.63

pora

0.62

Activations Density 0.671%

words related to actions or activities, especially those related to publication, premieres, releases, or revelations

past tense verbs and phrases indicating actions or events

No Comments

No Known Activations