INDEX

Explanations

phrases indicating completion or finality

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ggles

-0.74

hement

-0.71

iatrics

-0.70

iously

-0.69

etheus

-0.69

aires

-0.68

angles

-0.66

onomic

-0.66

kefeller

-0.66

atively

-0.65

POSITIVE LOGITS

 nutshell

0.87

 goodbye

0.79

 understatement

0.78

.ãĢį

0.76

 folks

0.75

 consolation

0.74

 description

0.72

0.71

SPONSORED

0.70

 alright

0.66

Activations Density 0.166%

phrases indicating completion or finality

No Comments

No Known Activations