INDEX

Explanations

entities or groups of people

oai_token-act-pair · gpt-3.5-turbo

terms related to groups of people or entities

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

tains

-0.88

 Contains

-0.74

 Causes

-0.71

 Adds

-0.71

 Says

-0.66

Adds

-0.66

 Allows

-0.65

 Needs

-0.64

needs

-0.63

 Does

-0.62

POSITIVE LOGITS

 were

1.60

 weren

1.55

 tended

1.51

 stayed

1.37

 wore

1.35

 flowed

1.34

 remained

1.33

 became

1.33

 retreated

1.32

were

1.31

Activations Density 0.470%

entities or groups of people

terms related to groups of people or entities

No Comments

No Known Activations