INDEX

Explanations

phrases related to items or actions that are left over or unused

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

haps

-0.65

paio

-0.65

ograph

-0.63

pport

-0.61

oto

-0.60

gging

-0.60

ocally

-0.59

king

-0.59

legraph

-0.58

ggle

-0.58

POSITIVE LOGITS

 leftover

0.96

 remnants

0.73

 residue

0.66

 chunks

0.62

 unused

0.60

 fallout

0.60

 ingredients

0.59

 memories

0.57

old

0.57

 scraps

0.57

Activations Density 11.744%

phrases related to items or actions that are left over or unused

No Comments

No Known Activations