INDEX

Explanations

hyphenated phrases and words

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

negative or critical sentiments in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

 caution

-0.74

 consolation

-0.68

 Hardy

-0.66

 Confederation

-0.64

 judgment

-0.64

civ

-0.63

 swirl

-0.62

 Territory

-0.61

 confidence

-0.61

 instant

-0.61

POSITIVE LOGITS

up

1.40

ups

1.26

upper

1.25

top

1.16

item

1.15

setting

1.15

feeding

1.09

smoking

1.09

piece

1.08

shift

1.07

Activations Density 0.052%

hyphenated phrases and words

negative or critical sentiments in the text

No Comments

No Known Activations