INDEX

Explanations

legal and political terms, particularly related to government bodies and actions

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

pload

-0.89

star

-0.71

 Citiz

-0.69

sha

-0.67

cop

-0.66

plet

-0.66

 glim

-0.66

nen

-0.66

atto

-0.63

INAL

-0.63

POSITIVE LOGITS

":["

0.69

 relating

0.68

ropy

0.64

("

0.62

(.

0.62

XIV

0.62

([

0.61

-----

0.60

:(

0.60

iferation

0.60

Activations Density 0.174%

legal and political terms, particularly related to government bodies and actions

No Comments

No Known Activations