INDEX

Explanations

phrases related to being up-to-date or current

oai_token-act-pair · gpt-3.5-turbo

instances of the word "up."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

VICE

-0.65

 CLSID

-0.55

 Frag

-0.53

 understatement

-0.51

uel

-0.51

 Prix

-0.50

aneously

-0.50

 Monster

-0.49

 disson

-0.49

POSITIVE LOGITS

rights

0.99

stairs

0.94

raised

0.94

votes

0.94

grading

0.93

swing

0.91

bringing

0.91

grades

0.90

bring

0.89

vote

0.89

Activations Density 0.068%

phrases related to being up-to-date or current

instances of the word "up."

No Comments

No Known Activations