INDEX

Explanations

dates expressed numerically with a slash '/' symbol indicating a range

oai_token-act-pair · gpt-3.5-turbo

the word "to" used 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 @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

pie

-0.63

ramer

-0.61

packing

-0.60

raped

-0.60

 downloaded

-0.60

 clicking

-0.59

wrote

-0.59

 Country

-0.58

 texted

-0.58

boarding

-0.57

POSITIVE LOGITS

 ensure

1.53

 avoid

1.51

 prevent

1.47

 appease

1.46

 minimize

1.45

 conserve

1.43

 accommodate

1.42

 facilitate

1.41

 maximize

1.40

 coincide

1.39

Activations Density 0.277%

dates expressed numerically with a slash '/' symbol indicating a range

the word "to" used in various contexts

No Comments

No Known Activations