INDEX

Explanations

mentions of academic or professional titles and affiliations

oai_token-act-pair · gpt-3.5-turbo

sentences or periods that mark the end of statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 glim

-0.90

 metic

-0.86

 defe

-0.84

 quir

-0.84

 prey

-0.81

 rall

-0.76

 overlooked

-0.74

 explan

-0.72

 emotion

-0.72

 footing

-0.72

POSITIVE LOGITS

 Additionally

1.07

 Likewise

1.02

 Also

1.01

 Similarly

1.01

 Meanwhile

1.01

 Later

1.01

 However

1.01

 Along

1.00

 Their

1.00

 Afterwards

0.95

Activations Density 1.410%

mentions of academic or professional titles and affiliations

sentences or periods that mark the end of statements

No Comments

No Known Activations