INDEX

Explanations

actions related to production or output, especially in the context of creating or generating large quantities

oai_token-act-pair · gpt-3.5-turbo

instances of the word "churn" and related forms

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 resemblance

-0.73

TERN

-0.67

 Target

-0.63

 Physicians

-0.61

 Telesc

-0.60

PRES

-0.60

cussion

-0.60

 Lennon

-0.59

 Defenders

-0.59

restrial

-0.59

POSITIVE LOGITS

 churn

1.25

ing

0.98

ingen

0.87

abal

0.84

zeb

0.83

ular

0.80

saw

0.79

edIn

0.78

ffield

0.77

ed

0.76

Activations Density 0.010%

actions related to production or output, especially in the context of creating or generating large quantities

instances of the word "churn" and related forms

No Comments

No Known Activations