INDEX

Explanations

phrases related to collecting and analyzing samples

oai_token-act-pair · gpt-3.5-turbo

references to samples in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

eve

-0.69

inance

-0.67

oggle

-0.67

capital

-0.67

chie

-0.64

 undone

-0.64

econom

-0.63

ESA

-0.63

occup

-0.63

sighted

-0.61

POSITIVE LOGITS

 samples

3.83

 sample

2.58

 sampled

2.02

 Sample

1.98

 sampling

1.95

sample

1.85

amples

1.84

Sample

1.82

 specimens

1.81

 demos

1.47

Activations Density 0.014%

phrases related to collecting and analyzing samples

references to samples in various contexts

No Comments

No Known Activations