INDEX

Explanations

references to physical objects

oai_token-act-pair · gpt-3.5-turbo

references to physical objects

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

BILITY

-0.83

 millenn

-0.77

ornia

-0.75

 Genie

-0.69

ther

-0.69

egal

-0.67

UTH

-0.66

ãĥ´

-0.64

ould

-0.64

sis

-0.63

POSITIVE LOGITS

 objects

0.95

objects

0.88

imus

0.81

ensity

0.79

*/(

0.73

 situated

0.71

 Objects

0.71

 silhou

0.70

ively

0.70

ivity

0.69

Activations Density 0.014%

references to physical objects

references to physical objects

No Comments

No Known Activations