INDEX

Explanations

words related to personal reflections, feelings of responsibility, and questioning aspects of society and culture

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-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

amps

-0.58

engineering

-0.56

ocular

-0.56

Reporting

-0.56

Plan

-0.55

 Measures

-0.55

amped

-0.54

sembly

-0.53

HQ

-0.53

 Romeo

-0.53

POSITIVE LOGITS

 gladly

1.09

 doubtless

1.08

 inevitably

1.07

 undoubtedly

1.05

 surely

1.02

be

1.00

 likely

0.96

 eventually

0.95

 probably

0.94

 never

0.90

Activations Density 13.477%

words related to personal reflections, feelings of responsibility, and questioning aspects of society and culture

No Comments

No Known Activations