INDEX

Explanations

mentions of the word "lower"

oai_token-act-pair · gpt-3.5-turbo

references to decreased levels or conditions

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

Pros

-0.80

POL

-0.77

vous

-0.74

Jew

-0.72

 Outbreak

-0.69

vp

-0.69

tnc

-0.68

Ze

-0.67

MO

-0.64

 Ramadan

-0.63

POSITIVE LOGITS

iating

1.01

 than

0.92

than

0.84

iation

0.84

case

0.84

 extrem

0.80

downs

0.79

down

0.77

iates

0.76

 pitched

0.76

Activations Density 0.016%

mentions of the word "lower"

references to decreased levels or conditions

No Comments

No Known Activations