INDEX

Explanations

objects or concepts that are the focus or target of actions or discussions

oai_token-act-pair · gpt-3.5-turbo

references to "objects" in various contexts

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

rosso

-0.76

fits

-0.75

illin

-0.73

hops

-0.70

breaks

-0.68

 Hayes

-0.67

HP

-0.65

 Surge

-0.65

 Merrill

-0.65

ews

-0.65

POSITIVE LOGITS

 object

3.79

 objects

2.71

object

2.61

 Object

2.42

 Objects

2.23

Object

2.09

objects

2.08

obj

1.39

Obj

1.27

 item

1.24

Activations Density 0.030%

objects or concepts that are the focus or target of actions or discussions

references to "objects" in various contexts

No Comments

No Known Activations