INDEX

Explanations

words related to continuations or follow-ups, such as "sequel", "prequel", "series", and "successor"

oai_token-act-pair · gpt-3.5-turbo

terms related to sequels, prequels, and tributes in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

lp

-0.59

_>

-0.58

uman

-0.57

isen

-0.56

cade

-0.54

bia

-0.53

gian

-0.53

asions

-0.52

Lens

-0.52

Query

-0.51

POSITIVE LOGITS

 thereto

1.54

to

1.43

To

0.99

to

0.98

TO

0.98

 unto

0.91

To

0.90

 towards

0.79

 toward

0.78

ta

0.78

Activations Density 0.331%

words related to continuations or follow-ups, such as "sequel", "prequel", "series", and "successor"

terms related to sequels, prequels, and tributes in various contexts

No Comments

No Known Activations