INDEX

Explanations

keywords related to variables and values in a program

oai_token-act-pair · gpt-3.5-turbo

references to values in a programming or technical context

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

spread

-0.76

kefeller

-0.76

ocamp

-0.74

mare

-0.73

ooks

-0.71

shaw

-0.71

dar

-0.70

STON

-0.70

ned

-0.69

nee

-0.68

POSITIVE LOGITS

 values

0.89

 value

0.88

values

0.76

EStream

0.76

value

0.73

 output

0.71

 Value

0.70

 proposition

0.69

 outputs

0.68

VALUE

0.67

Activations Density 0.014%

keywords related to variables and values in a program

references to values in a programming or technical context

No Comments

No Known Activations