INDEX

Explanations

mentions of the word "bumble" with varying intensity

oai_token-act-pair · gpt-3.5-turbo

instances of the word "bumble."

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ablishment

-0.86

ohn

-0.82

hips

-0.75

olphins

-0.73

atural

-0.73

hetical

-0.72

thening

-0.69

aca

-0.67

aunder

-0.67

afort

-0.66

POSITIVE LOGITS

weed

1.10

BLE

0.78

BUG

0.77

umble

0.77

Redd

0.76

ministic

0.75

grass

0.74

bee

0.72

 Britann

0.72

 Doodle

0.71

Activations Density 0.007%

mentions of the word "bumble" with varying intensity

instances of the word "bumble."

No Known Activations