INDEX

Explanations

instances of the word "many" and its associated numbers

oai_token-act-pair · gpt-3.5-turbo

phrases that question or quantify the number of items or occurrences

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

avior

-0.84

ESE

-0.79

afort

-0.75

andise

-0.75

¶æ

-0.74

vertisement

-0.72

mast

-0.70

igation

-0.70

aviour

-0.70

amphetamine

-0.70

POSITIVE LOGITS

 times

1.20

 mouths

1.02

 iterations

0.99

 layers

0.99

 hours

0.97

 calories

0.96

 cooks

0.93

 variables

0.92

 thousand

0.91

 facets

0.91

Activations Density 0.068%

instances of the word "many" and its associated numbers

phrases that question or quantify the number of items or occurrences

No Comments

No Known Activations