INDEX

Explanations

words related to health conditions and healthcare

oai_token-act-pair · gpt-3.5-turbo

references to health-related subjects and discussions

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

 throats

-0.72

 awhile

-0.70

 square

-0.69

 Cree

-0.69

 piles

-0.67

 throat

-0.66

 Morrow

-0.66

 Jail

-0.65

 sockets

-0.65

 Carth

-0.65

POSITIVE LOGITS

related

1.76

themed

1.61

focused

1.56

specific

1.49

sensitive

1.46

oriented

1.44

friendly

1.43

based

1.43

inspired

1.40

centered

1.35

Activations Density 0.066%

words related to health conditions and healthcare

references to health-related subjects and discussions

No Comments

No Known Activations