INDEX

Explanations

specific patterns of characters, potentially related to a particular language or type of text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Cosponsors

-0.67

adelphia

-0.64

 Mehran

-0.63

 ratified

-0.61

agall

-0.60

*/(

-0.60

onut

-0.58

 willingly

-0.58

lication

-0.58

issance

-0.56

POSITIVE LOGITS

culosis

0.68

Ö

0.64

Ø¯

0.60

ibaba

0.60

 lithium

0.58

zbek

0.58

¥µ

0.57

VALUE

0.57

iosity

0.55

schild

0.54

Activations Density 13.455%

specific patterns of characters, potentially related to a particular language or type of text

No Comments

No Known Activations