INDEX

Explanations

the phrase "conflicts of interest."

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

LESS

-0.76

nz

-0.70

tv

-0.65

FN

-0.64

RAY

-0.63

 Jelly

-0.63

 starve

-0.62

©¶æ¥µ

-0.61

uristic

-0.61

 Lauder

-0.60

POSITIVE LOGITS

 interests

0.79

interest

0.77

 Interest

0.76

 interest

0.74

rontal

0.69

ego

0.69

 tongues

0.65

 sorts

0.65

 wills

0.64

Interest

0.63

Activations Density 0.073%

the phrase "conflicts of interest."

No Comments

No Known Activations