INDEX

Explanations

phrases related to levels of achievement or proficiency

oai_token-act-pair · gpt-3.5-turbo

references to academic degrees and their significance

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

aways

-0.76

istar

-0.74

wal

-0.73

gets

-0.72

ages

-0.71

backs

-0.69

 Junk

-0.63

back

-0.63

owned

-0.62

stock

-0.62

POSITIVE LOGITS

 degree

3.88

 Degree

2.84

degree

2.72

 degrees

2.61

 extent

2.08

Deg

1.66

 magnitude

1.46

deg

1.43

Âº

1.41

Â°

1.41

Activations Density 0.004%

phrases related to levels of achievement or proficiency

references to academic degrees and their significance

No Comments

No Known Activations