INDEX

Explanations

words related to political events, actions, and controversies

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

enhagen

-0.83

swick

-0.83

abase

-0.75

eday

-0.72

LESS

-0.72

HC

-0.71

Bank

-0.71

iths

-0.70

RESULTS

-0.69

IFIED

-0.68

POSITIVE LOGITS

 measures

1.00

 crus

0.95

 sentiment

0.95

 Measures

0.91

 campaigner

0.90

 campaigners

0.83

 deterrent

0.83

 efforts

0.82

 initiatives

0.81

 antibodies

0.80

Activations Density 0.063%

words related to political events, actions, and controversies

No Comments

No Known Activations