INDEX

Explanations

the prefix "di" followed by a 9 and 10-letter word

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "di."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãĤ¤ãĥĪ

-0.87

éĹĺ

-0.83

INGS

-0.81

manship

-0.79

ãģĨ

-0.79

Thumbnail

-0.77

 Seah

-0.74

Reviewer

-0.72

ilater

-0.72

ORTS

-0.71

POSITIVE LOGITS

abol

1.25

ocese

1.17

abolic

1.16

urnal

1.16

agon

1.05

ablo

0.94

aper

0.92

utsch

0.90

agram

0.90

agonal

0.87

Activations Density 0.007%

the prefix "di" followed by a 9 and 10-letter word

occurrences of the prefix "di."

No Comments

No Known Activations