INDEX

Explanations

acronyms related to government agencies or programs

oai_token-act-pair · gpt-3.5-turbo

references to various military and defense-related programs or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ĸļ

-0.93

 Subtle

-0.87

 Tosh

-0.74

itars

-0.73

ãĥ¼ãĥĨãĤ£

-0.72

 Sorceress

-0.71

 Healer

-0.71

 Swordsman

-0.71

ãĤ´ãĥ³

-0.70

 Haku

-0.70

POSITIVE LOGITS

SA

1.08

ISA

1.08

CP

1.07

SE

1.06

NP

1.04

WP

1.04

CS

1.02

HS

1.01

GC

1.00

PA

0.99

Activations Density 0.309%

acronyms related to government agencies or programs

references to various military and defense-related programs or entities

No Comments

No Known Activations