INDEX

Explanations

phrases related to completing, finishing, or summarizing something

oai_token-act-pair · gpt-3.5-turbo

phrases related to concluding or finishing actions

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

ogi

-0.73

acus

-0.72

 Haas

-0.68

appropri

-0.62

 alter

-0.61

bee

-0.60

caf

-0.59

acist

-0.58

 spoof

-0.57

chens

-0.56

POSITIVE LOGITS

Ń·

0.91

sett

0.78

 Winds

0.72

¥µ

0.69

EVA

0.67

lished

0.66

mortem

0.65

 landfall

0.65

byss

0.65

doors

0.65

Activations Density 0.054%

phrases related to completing, finishing, or summarizing something

phrases related to concluding or finishing actions

No Comments

No Known Activations