INDEX

Explanations

terms related to the timing of events in a sequence

oai_token-act-pair · gpt-3.5-turbo

temporal references and indicators related to time

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

onom

-0.66

oret

-0.63

actic

-0.61

ho

-0.56

correct

-0.56

hal

-0.56

]'

-0.55

Oregon

-0.54

PAN

-0.54

PAC

-0.54

POSITIVE LOGITS

*/(

0.64

veyard

0.59

ĸļ

0.58

illac

0.58

iatrics

0.56

isine

0.56

igers

0.55

 ratification

0.55

eatures

0.55

widget

0.54

Activations Density 0.719%

terms related to the timing of events in a sequence

temporal references and indicators related to time

No Comments

No Known Activations