INDEX

Explanations

phrases indicating a strong suggestion or clue

oai_token-act-pair · gpt-3.5-turbo

phrases related to indicators or signs of trends 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

hibited

-0.70

morph

-0.66

tz

-0.66

married

-0.60

oos

-0.60

space

-0.58

onyms

-0.58

 practiced

-0.57

osh

-0.57

yright

-0.57

POSITIVE LOGITS

 indication

1.01

 hint

0.89

 hints

0.82

 indications

0.81

 Signs

0.78

 glim

0.78

 clue

0.76

 signs

0.75

 signal

0.75

 glimpse

0.75

Activations Density 0.086%

phrases indicating a strong suggestion or clue

phrases related to indicators or signs of trends or events

No Comments

No Known Activations