INDEX

Explanations

phrases indicating something is considered or seen in a certain way

oai_token-act-pair · gpt-3.5-turbo

phrases that denote perceptions or evaluations of something as being significant or noteworthy

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

 Sieg

-0.72

ammy

-0.70

 takeoff

-0.66

apeake

-0.66

 rain

-0.65

driving

-0.63

zzy

-0.62

mith

-0.61

 riff

-0.61

 hammer

-0.60

POSITIVE LOGITS

enance

1.00

phas

0.87

CLASSIFIED

0.80

æĦ

0.80

æĺ¯

0.78

MFT

0.78

ä»£

0.76

åº

0.75

sburg

0.73

recated

0.73

Activations Density 0.024%

phrases indicating something is considered or seen in a certain way

phrases that denote perceptions or evaluations of something as being significant or noteworthy

No Comments

No Known Activations