INDEX

Explanations

mentions of how something is organized or structured

oai_token-act-pair · gpt-3.5-turbo

actions and how they are organized or structured in various contexts

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

OTOS

-0.66

voy

-0.65

ATK

-0.63

 Boat

-0.62

oids

-0.59

 âĨ

-0.59

clave

-0.59

sonian

-0.59

owners

-0.58

TERN

-0.57

POSITIVE LOGITS

 behave

0.79

 unfolded

0.76

 fared

0.75

ribed

0.71

 structured

0.69

 differs

0.69

 mang

0.68

 reacted

0.66

 proced

0.65

 fare

0.65

Activations Density 0.372%

mentions of how something is organized or structured

actions and how they are organized or structured in various contexts

No Comments

No Known Activations