INDEX

Explanations

variations or changes

oai_token-act-pair · gpt-3.5-turbo

terms related to variation and variability

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

 financially

-0.75

ofi

-0.73

Pand

-0.72

Â·Â·

-0.71

need

-0.69

ocket

-0.69

bags

-0.69

iar

-0.65

 Hoover

-0.64

NSA

-0.62

POSITIVE LOGITS

 variation

3.76

 variations

3.15

 variability

2.59

 Vari

2.20

 variance

2.10

Vari

1.95

 variants

1.88

 variant

1.78

 fluctuations

1.69

 deviations

1.61

Activations Density 0.016%

variations or changes

terms related to variation and variability

No Comments

No Known Activations