INDEX

Explanations

words related to ability and possibility

oai_token-act-pair · gpt-3.5-turbo

conjunctions that connect conditions or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Gleaming

-0.71

okÃ©

-0.67

ainted

-0.65

 Raider

-0.64

iHUD

-0.63

CRIP

-0.61

 Subtle

-0.61

advertisement

-0.60

ogie

-0.59

asher

-0.59

POSITIVE LOGITS

 SHOULD

1.04

 should

1.03

 ought

1.01

should

1.00

 must

0.96

 shouldn

0.92

 cannot

0.91

must

0.86

 wont

0.85

 will

0.85

Activations Density 0.068%

words related to ability and possibility

conjunctions that connect conditions or actions

No Comments

No Known Activations