INDEX

Explanations

mentions of the name 'Ins'

oai_token-act-pair · gpt-3.5-turbo

patterns related to the suffix "ins."

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

¶ħ

-0.68

ŃĶ

-0.67

¿½

-0.66

 severed

-0.65

 divest

-0.65

 sever

-0.64

Ner

-0.64

 taxpayer

-0.64

 Nanto

-0.62

 campuses

-0.61

POSITIVE LOGITS

pection

1.17

urance

1.05

ufficient

1.03

anity

1.02

pect

1.02

ins

1.02

urg

1.00

piration

0.98

urgical

0.94

terness

0.94

Activations Density 0.012%

mentions of the name 'Ins'

patterns related to the suffix "ins."

No Comments

No Known Activations