INDEX

Explanations

product advertisements and health-related information

oai_token-act-pair · gpt-3.5-turbo

elements related to advertisements and promotional content

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

)).

-0.42

etc

-0.40

}.

-0.40

 respectively

-0.39

 fame

-0.38

).

-0.35

SPONSORED

-0.35

%).

-0.35

".

-0.34

 vicinity

-0.34

POSITIVE LOGITS

 however

0.43

 Explan

0.39

 meanwhile

0.38

xtap

0.38

 depends

0.37

ogether

0.37

 concedes

0.36

 caveats

0.36

arij

0.36

 glim

0.36

Activations Density 5.346%

product advertisements and health-related information

elements related to advertisements and promotional content

No Comments

No Known Activations