INDEX

Explanations

the word "stacked" or other related terms

oai_token-act-pair · gpt-3.5-turbo

terms related to accumulation or grouping items together

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

-0.87

hood

-0.80

bler

-0.79

zie

-0.77

cot

-0.76

context

-0.76

nel

-0.72

kay

-0.72

lde

-0.72

Poo

-0.70

POSITIVE LOGITS

 stacked

1.38

 stacking

1.08

 millenn

1.05

 stacks

0.97

 unbeliev

0.91

 adolesc

0.90

ikuman

0.84

 arrays

0.83

 shenan

0.83

 brackets

0.83

Activations Density 0.006%

the word "stacked" or other related terms

terms related to accumulation or grouping items together

No Comments

No Known Activations