INDEX

Explanations

words related to leadership or authority, particularly those involving the concept of 'presiding'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "preside" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ISM

-0.79

 Tigers

-0.68

rosis

-0.67

 Panther

-0.66

beans

-0.66

 Paradise

-0.65

orne

-0.64

riad

-0.64

Bear

-0.63

 Rite

-0.63

POSITIVE LOGITS

 pres

1.27

ervative

1.04

ervatives

1.02

orial

0.85

leep

0.82

Pres

0.81

umes

0.80

byter

0.79

 conduc

0.79

 perspect

0.76

Activations Density 0.006%

words related to leadership or authority, particularly those involving the concept of 'presiding'

occurrences of the word "preside" and its variations

No Comments

No Known Activations