INDEX

Explanations

phrases indicating a state of crisis or conflict

oai_token-act-pair · gpt-3.5-turbo

the word "in" indicating location or context

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

 VIDEOS

-0.70

»Ĵ

-0.68

SHIP

-0.67

yles

-0.66

NOW

-0.65

uces

-0.64

 Sorceress

-0.62

soever

-0.61

 portrays

-0.61

 abuses

-0.61

POSITIVE LOGITS

escap

1.23

ordinate

1.15

 accordance

1.14

animate

1.10

bred

1.09

operative

1.07

clus

1.05

visible

1.05

 lieu

1.04

humane

1.03

Activations Density 0.178%

phrases indicating a state of crisis or conflict

the word "in" indicating location or context

No Comments

No Known Activations