INDEX

Explanations

phrases related to correctness or moral judgment

oai_token-act-pair · gpt-3.5-turbo

phrases that affirm correctness or appropriateness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

mat

-0.68

ĸļ

-0.64

leneck

-0.64

ulative

-0.62

ains

-0.61

arette

-0.60

ature

-0.58

oute

-0.57

chio

-0.57

aden

-0.57

POSITIVE LOGITS

eous

1.13

 wing

0.77

wing

0.76

move

0.70

shore

0.70

terday

0.68

 winger

0.68

 forearm

0.68

ward

0.68

å¾

0.67

Activations Density 0.054%

phrases related to correctness or moral judgment

phrases that affirm correctness or appropriateness

No Comments

No Known Activations