INDEX

Explanations

words related to conflict, war, and fighting

oai_token-act-pair · gpt-3.5-turbo

terms related to war and conflict situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Chance

-0.77

 Swim

-0.73

racuse

-0.73

ministic

-0.72

verett

-0.69

 Tennis

-0.68

DMV

-0.67

é¾įå

-0.66

 typew

-0.66

NX

-0.65

POSITIVE LOGITS

 raged

1.16

 waged

1.13

 raging

1.09

torn

1.06

lord

1.01

lords

1.00

 ravaged

0.96

 erupted

0.95

 engulf

0.92

 insurgents

0.89

Activations Density 0.113%

words related to conflict, war, and fighting

terms related to war and conflict situations

No Comments

No Known Activations