INDEX

Explanations

phrases related to geographical locations or terms

oai_token-act-pair · gpt-3.5-turbo

references to the word "strat" 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 @ 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

fulness

-0.68

 Referred

-0.68

manship

-0.65

give

-0.65

GRE

-0.64

hours

-0.64

Beast

-0.63

Ever

-0.63

 Recall

-0.63

ternity

-0.63

POSITIVE LOGITS

 strat

1.20

osp

0.96

itud

0.93

raph

0.87

ixtape

0.83

imore

0.83

igm

0.81

ategy

0.81

itudes

0.80

imer

0.79

Activations Density 0.006%

phrases related to geographical locations or terms

references to the word "strat" in various contexts

No Comments

No Known Activations