INDEX

Explanations

specific references to something, possibly objects or actions

oai_token-act-pair · gpt-3.5-turbo

references to the word "those" 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ob

-0.77

ILY

-0.73

Â¨

-0.71

iness

-0.71

onis

-0.70

achus

-0.69

BW

-0.67

maker

-0.67

shapeshifter

-0.67

manship

-0.66

POSITIVE LOGITS

 pesky

1.10

 kinds

0.94

 sorts

0.86

 fateful

0.81

 damned

0.74

 sights

0.73

 nifty

0.73

 aforementioned

0.73

 darn

0.70

 same

0.69

Activations Density 0.066%

specific references to something, possibly objects or actions

references to the word "those" in various contexts

No Comments

No Known Activations