INDEX

Explanations

personal opinions and beliefs expressed with emotive language

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 littered

-0.60

 Actions

-0.59

CI

-0.58

senal

-0.57

 bount

-0.57

assault

-0.57

CI

-0.56

 Rounds

-0.56

 Flags

-0.55

rant

-0.55

POSITIVE LOGITS

 afford

1.40

 muster

1.08

 feas

1.05

 conceive

0.97

athom

0.96

 cope

0.95

 withstand

0.95

 decipher

0.94

 imagine

0.94

 stomach

0.93

Activations Density 1.732%

personal opinions and beliefs expressed with emotive language

No Comments

No Known Activations