INDEX

Explanations

long-range weapon-related terms

oai_token-act-pair · gpt-3.5-turbo

terms related to missile range and capabilities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 monet

-0.74

 wallpaper

-0.68

 idol

-0.65

 approval

-0.63

 perpet

-0.63

 Millenn

-0.63

 shutting

-0.62

 evangel

-0.61

 OFFIC

-0.61

 sponsored

-0.61

POSITIVE LOGITS

range

4.46

ranged

2.51

Range

2.10

ranging

2.04

 range

1.99

rang

1.96

 Range

1.78

distance

1.73

 ranges

1.56

 ranged

1.18

Activations Density 0.016%

long-range weapon-related terms

terms related to missile range and capabilities

No Comments

No Known Activations