INDEX

Explanations

places or instances where the word "note" is used

oai_token-act-pair · gpt-3.5-turbo

phrases indicating conditionality or the presence of specific criteria

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

endez

-0.53

qt

-0.48

anger

-0.47

Els

-0.46

unal

-0.45

anamo

-0.45

pler

-0.43

venge

-0.42

emies

-0.41

opian

-0.41

POSITIVE LOGITS

 comprom

0.47

 overriding

0.44

 percentage

0.42

 totality

0.42

 construed

0.40

 fortun

0.40

 temptation

0.39

 occurrences

0.39

"%

0.39

 prolonged

0.39

Activations Density 0.894%

places or instances where the word "note" is used

phrases indicating conditionality or the presence of specific criteria

No Comments

No Known Activations