INDEX

Explanations

instances where someone or something is getting a hold of something or someone, often emphasizing exerting control or influence over a situation

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-ajt

Prompts (Dashboard)

12,288 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

46,080

Data Type

torch.float32

Hook Point

blocks.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

apollo-research/Skylion007-openwebtext-tokenizer-gpt2

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Audit

-0.89

 Chou

-0.74

ophile

-0.73

 starved

-0.71

RFC

-0.71

 submar

-0.70

 Tsukuyomi

-0.70

anchester

-0.70

 extortion

-0.69

osal

-0.68

POSITIVE LOGITS

hold

1.54

holding

1.19

holder

1.18

holders

1.07

erness

1.04

fast

1.04

ings

1.04

yer

1.01

backs

0.99

pipe

0.97

Activations Density 1.301%

instances where someone or something is getting a hold of something or someone, often emphasizing exerting control or influence over a situation

No Comments

No Known Activations