INDEX

Explanations

references to a central or foundational element

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "base."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

icut

-0.74

ippi

-0.69

selves

-0.69

inness

-0.68

NESS

-0.67

odd

-0.67

byter

-0.67

andro

-0.66

apes

-0.65

tnc

-0.64

POSITIVE LOGITS

camp

0.93

 base

0.84

plate

0.79

plan

0.74

base

0.72

plates

0.72

 camp

0.70

 structure

0.70

 erosion

0.69

ball

0.68

Activations Density 0.012%

references to a central or foundational element

repeated mentions of the word "base."

No Comments

No Known Activations