INDEX

Explanations

terms related to adversaries or opponents

oai_token-act-pair · gpt-3.5-turbo

references to adversarial or opposing groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

eret

-0.74

ascript

-0.74

authorized

-0.73

oled

-0.73

printed

-0.72

razil

-0.71

Ħ¢

-0.70

estone

-0.69

otto

-0.69

arrell

-0.68

POSITIVE LOGITS

 Enemy

1.02

foe

0.91

 enemy

0.86

 combatants

0.82

 vanquished

0.77

 hated

0.75

 legion

0.75

 Enemies

0.74

emies

0.74

 waging

0.73

Activations Density 0.025%

terms related to adversaries or opponents

references to adversarial or opposing groups

No Comments

No Known Activations