INDEX

Explanations

words related to protection or defense, particularly the word "shield."

oai_token-act-pair · gpt-3.5-turbo

terms related to protective barriers or defenses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

cause

-0.76

gres

-0.76

iasco

-0.75

orie

-0.74

 Helpful

-0.73

 âĸº

-0.73

eph

-0.71

pheus

-0.70

 Judicial

-0.69

orio

-0.68

POSITIVE LOGITS

 shields

1.00

 shield

0.99

maid

0.86

 shielding

0.86

heed

0.83

igans

0.79

Shield

0.78

shield

0.76

 piercing

0.76

 curtain

0.74

Activations Density 0.006%

words related to protection or defense, particularly the word "shield."

terms related to protective barriers or defenses

No Comments

No Known Activations