INDEX

Explanations

verbs related to defense or protection

oai_token-act-pair · gpt-3.5-turbo

terms related to defense and neglect

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Jarrett

-0.81

anamo

-0.73

itable

-0.70

tyr

-0.67

 stakes

-0.65

 bonds

-0.65

raltar

-0.64

Madison

-0.61

 teammate

-0.61

 Survivors

-0.61

POSITIVE LOGITS

Ĳ

0.92

Ķ

0.88

rodu

0.87

 fend

0.82

ĳ

0.80

iard

0.80

ual

0.79

)=(

0.79

eenth

0.77

abouts

0.76

Activations Density 0.024%

verbs related to defense or protection

terms related to defense and neglect

No Comments

No Known Activations