INDEX

Explanations

phrases related to legal rulings, events, and governmental operations

oai_token-act-pair · gpt-3.5-turbo

references to significant political events or judicial rulings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Favorite

-0.68

 Learns

-0.66

natureconservancy

-0.61

 anecd

-0.59

 Trainer

-0.58

maxwell

-0.58

depending

-0.57

 Profile

-0.57

 ILCS

-0.57

 Trails

-0.55

POSITIVE LOGITS

 sacked

0.75

 abolished

0.74

 reopened

0.74

 reinstated

0.73

 dumped

0.72

 inaug

0.71

 unilaterally

0.71

 outlawed

0.70

 enshr

0.70

 privat

0.70

Activations Density 1.856%

phrases related to legal rulings, events, and governmental operations

references to significant political events or judicial rulings

No Comments

No Known Activations