INDEX

Explanations

numeric series in the format "Number Number" referring to years and their durations

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

ularity

-0.64

tre

-0.64

overty

-0.63

lance

-0.61

piring

-0.61

alam

-0.61

auga

-0.60

ogh

-0.60

qqa

-0.59

alling

-0.58

POSITIVE LOGITS

 fraction

0.65

 authenticated

0.64

 survives

0.64

inally

0.63

ieth

0.63

 lasts

0.58

acebook

0.55

can

0.55

can

0.55

 exceptions

0.54

Activations Density 18.194%

numeric series in the format "Number Number" referring to years and their durations

No Comments

No Known Activations