INDEX

Explanations

terms related to technical specifications or classifications, specifically related to different types

oai_token-act-pair · gpt-3.5-turbo

references to specific product types or classifications

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aud

-0.77

sterdam

-0.73

jud

-0.71

 prevail

-0.68

 meantime

-0.66

 liberties

-0.65

 camer

-0.64

adows

-0.64

 anthem

-0.63

terday

-0.63

POSITIVE LOGITS

 Type

3.72

Type

2.38

 TYPE

2.10

 type

1.98

 Types

1.98

type

1.78

TYPE

1.60

Typ

1.55

 types

1.50

types

1.43

Activations Density 0.012%

terms related to technical specifications or classifications, specifically related to different types

references to specific product types or classifications

No Comments

No Known Activations