INDEX

Explanations

mentions of or references to swords

oai_token-act-pair · gpt-3.5-turbo

references to swords

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

-0.78

SPONSORED

-0.76

Newsletter

-0.73

greSQL

-0.71

Organ

-0.71

ipolar

-0.69

EMBER

-0.68

-0.67

Interstitial

-0.67

ODE

-0.67

POSITIVE LOGITS

 swords

1.01

 swall

1.00

fish

0.99

 blades

0.93

 sword

0.92

breaker

0.91

fights

0.89

fighting

0.87

 wielding

0.86

fighter

0.85

Activations Density 0.014%

mentions of or references to swords

references to swords

No Comments

No Known Activations