INDEX

Explanations

phrases related to criticism or concern about a particular person's actions or behavior

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

Negative Logits

ãĤ¹ãĥĪ

-0.86

Ĭ±

-0.81

icons

-0.78

acers

-0.77

ATURES

-0.76

esm

-0.76

unctions

-0.76

aws

-0.75

quarters

-0.74

ready

-0.73

POSITIVE LOGITS

guy

1.15

 sucker

0.98

 stuff

0.94

 particular

0.93

 week

0.92

 thing

0.92

 weekend

0.90

 kind

0.88

 country

0.88

 situation

0.88

Activations Density 0.176%

phrases related to criticism or concern about a particular person's actions or behavior

No Known Activations