INDEX

Explanations

words starting with 'a' that are likely part of philosophical or contemplative text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the indefinite article 'a' across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

NPR

-0.83

advertising

-0.80

 Labs

-0.78

Austin

-0.77

kids

-0.76

Apps

-0.75

 Emails

-0.74

apps

-0.74

 ESPN

-0.74

 Amtrak

-0.74

POSITIVE LOGITS

ught

1.09

 pious

1.08

rouse

1.07

 pecul

1.04

 magnificent

1.04

 remnant

1.03

 definite

1.02

 servant

1.01

 manifestation

1.01

 noble

1.00

Activations Density 0.303%

words starting with 'a' that are likely part of philosophical or contemplative text

occurrences of the indefinite article 'a' across various contexts

No Comments

No Known Activations