INDEX

Explanations

words related to sequences of events or actions

oai_token-act-pair · gpt-3.5-turbo

terms related to sequential events or consequences

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

 relax

-0.64

mx

-0.60

bos

-0.58

warm

-0.58

bones

-0.58

uras

-0.57

bone

-0.56

 Flags

-0.56

wings

-0.56

 Write

-0.56

POSITIVE LOGITS

 thereafter

3.09

 ensuing

3.01

 subsequent

2.27

 ensued

1.70

sequent

1.61

sequently

1.60

 onward

1.55

 afterward

1.53

 resultant

1.48

 consequ

1.46

Activations Density 0.021%

words related to sequences of events or actions

terms related to sequential events or consequences

No Comments

No Known Activations