INDEX

Explanations

references to things or concepts being familiar or recognizable

oai_token-act-pair · gpt-3.5-turbo

phrases indicating recognition or familiarity

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

adal

-0.68

atown

-0.68

hers

-0.67

wings

-0.66

ritional

-0.66

mental

-0.64

 reins

-0.64

onding

-0.64

ascript

-0.63

Display

-0.63

POSITIVE LOGITS

 sounding

0.86

 alarms

0.83

 bells

0.77

 drums

0.73

 Echo

0.67

sounding

0.67

 Morse

0.66

 trumpet

0.66

wegian

0.66

 aloud

0.66

Activations Density 0.087%

references to things or concepts being familiar or recognizable

phrases indicating recognition or familiarity

No Comments

No Known Activations