INDEX

Explanations

terms related to providing commentary or analysis

oai_token-act-pair · gpt-3.5-turbo

references to commentary and analysis in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

riages

-0.74

bis

-0.70

rir

-0.66

 Lans

-0.64

wikipedia

-0.64

chen

-0.63

nesses

-0.63

 Dull

-0.62

bie

-0.62

 Eucl

-0.61

POSITIVE LOGITS

 commentary

0.95

mund

0.78

jad

0.77

ature

0.72

spective

0.68

isf

0.68

 commentator

0.67

 cartoons

0.67

raint

0.66

atures

0.66

Activations Density 0.023%

terms related to providing commentary or analysis

references to commentary and analysis in various contexts

No Comments

No Known Activations