INDEX

Explanations

phrases related to organizing information or tasks

oai_token-act-pair · gpt-3.5-turbo

possessive pronouns indicating ownership or reference to personal items

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

trak

-0.89

 Uriel

-0.80

 Originally

-0.78

apo

-0.77

Lago

-0.75

ilts

-0.73

wik

-0.72

 Goes

-0.71

raq

-0.68

vous

-0.67

POSITIVE LOGITS

own

1.66

 favourite

1.08

 favorite

1.05

 opponent

1.04

ocard

1.00

 adversary

0.99

 desired

0.98

 imagination

0.94

selves

0.93

 entire

0.93

Activations Density 0.121%

phrases related to organizing information or tasks

possessive pronouns indicating ownership or reference to personal items

No Comments

No Known Activations