INDEX

Explanations

mentions or references to military or combat-related terms and concepts

oai_token-act-pair · gpt-3.5-turbo

references to military organizations or combat forces

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Increased

-0.73

venge

-0.69

 Previously

-0.67

urther

-0.66

 Reverse

-0.66

oward

-0.66

 benefiting

-0.62

 reversing

-0.62

 renewed

-0.59

 Improved

-0.59

POSITIVE LOGITS

 barely

1.13

 scarcely

1.12

 hardly

1.07

only

1.03

 nowhere

1.02

 seldom

1.01

 rarely

1.01

 merely

0.98

 mere

0.96

 only

0.95

Activations Density 0.937%

mentions or references to military or combat-related terms and concepts

references to military organizations or combat forces

No Comments

No Known Activations