INDEX

Explanations

personal pronouns combined with verbs that imply a result or outcome

oai_token-act-pair · gpt-3.5-turbo

references to personal connections and interactions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Accessory

-0.64

AMI

-0.61

 Bare

-0.60

 dictates

-0.60

 Mandarin

-0.59

fact

-0.59

ami

-0.59

Effect

-0.58

 Firm

-0.57

common

-0.57

POSITIVE LOGITS

 hooked

1.02

 acquainted

1.01

 addicted

0.94

 pumped

0.81

 traction

0.81

 excited

0.78

 booted

0.77

 accustomed

0.76

 stuck

0.76

 drunk

0.76

Activations Density 0.091%

personal pronouns combined with verbs that imply a result or outcome

references to personal connections and interactions

No Comments

No Known Activations