INDEX

Explanations

words related to permission or authorization

oai_token-act-pair · gpt-3.5-turbo

the phrase "wouldn't" and its variations

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

ocal

-0.68

ARM

-0.67

æ³

-0.65

core

-0.65

story

-0.63

 Case

-0.63

Dialog

-0.63

 Gong

-0.61

agency

-0.61

dress

-0.60

POSITIVE LOGITS

't

1.19

 never

0.84

 proble

0.84

 surely

0.82

terness

0.77

ÃĥÃĤ

0.76

 hardly

0.76

 tremend

0.76

 adjourn

0.75

itiveness

0.75

Activations Density 0.011%

words related to permission or authorization

the phrase "wouldn't" and its variations

No Comments

No Known Activations