INDEX

Explanations

verbs ending in the letter 'n' with high activations

oai_token-act-pair · gpt-3.5-turbo

words related to legal or bureaucratic terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

BMC

-0.76

KP

-0.73

Cot

-0.72

Bee

-0.69

 Mayer

-0.69

KC

-0.67

 Supervisor

-0.66

 bonds

-0.66

 liaison

-0.64

 Baxter

-0.63

POSITIVE LOGITS

actly

1.29

theless

1.15

terday

1.10

vernment

1.05

imately

1.04

respective

1.04

undreds

1.03

ogether

1.03

pecially

1.03

ombies

1.00

Activations Density 0.113%

verbs ending in the letter 'n' with high activations

words related to legal or bureaucratic terminology

No Comments

No Known Activations