INDEX

Explanations

words related to military combat tactics and strategies, especially those involving interacting with enemies

oai_token-act-pair · gpt-3.5-turbo

references to opponents or adversaries within a competitive context

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

uesday

-0.87

ocument

-0.80

ascript

-0.79

oled

-0.78

hetti

-0.77

mbuds

-0.77

aus

-0.76

phrine

-0.74

aza

-0.73

ranging

-0.73

POSITIVE LOGITS

 combatants

1.04

 enemy

0.96

 Enemy

0.94

 takeover

0.88

 invasion

0.82

enemy

0.78

foe

0.78

 spouse

0.78

 commander

0.77

 fighters

0.77

Activations Density 0.020%

words related to military combat tactics and strategies, especially those involving interacting with enemies

references to opponents or adversaries within a competitive context

No Comments

No Known Activations