INDEX

Explanations

numerical references in the form of the word "aph."

oai_token-act-pair · gpt-3.5-turbo

occurrences of "aph" in various contexts and formats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

PUT

-0.90

nings

-0.80

âĸ¬âĸ¬

-0.74

swick

-0.74

 Supplemental

-0.70

 Borderlands

-0.70

 fitting

-0.67

LOAD

-0.66

éĹĺ

-0.64

 Kaepernick

-0.63

POSITIVE LOGITS

obia

1.21

ysics

1.16

aph

1.12

ysical

1.10

osate

1.06

orically

1.06

armac

1.03

alon

1.03

obic

1.02

aith

0.98

Activations Density 0.010%

numerical references in the form of the word "aph."

occurrences of "aph" in various contexts and formats

No Comments

No Known Activations