INDEX

Explanations

specific number phrases and keywords related to various topics

oai_token-act-pair · gpt-3.5-turbo

repetitive conjunctions or phrases indicating addition or continuity

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

Deal

-0.68

qt

-0.64

icc

-0.63

BIP

-0.62

imus

-0.62

brush

-0.61

spir

-0.61

Ĭ±

-0.61

FINE

-0.60

LinkedIn

-0.60

POSITIVE LOGITS

 secondly

1.00

romeda

0.86

rew

0.83

ersen

0.82

 THEN

0.79

 furthermore

0.78

hra

0.78

rogens

0.74

 then

0.72

alus

0.70

Activations Density 0.121%

specific number phrases and keywords related to various topics

repetitive conjunctions or phrases indicating addition or continuity

No Comments

No Known Activations