INDEX

Explanations

references to things being at the center of attention or focus

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a central or main focus in a context

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ishable

-0.76

utan

-0.64

à©

-0.64

é¾įå

-0.61

)--

-0.60

hua

-0.60

syn

-0.60

 à¨

-0.60

ratom

-0.59

 chops

-0.59

POSITIVE LOGITS

Initialized

0.78

 gravity

0.77

igm

0.70

rency

0.69

inence

0.69

ierre

0.67

EVA

0.67

pole

0.66

gie

0.65

dyl

0.63

Activations Density 0.067%

references to things being at the center of attention or focus

phrases that indicate a central or main focus in a context

No Known Activations