INDEX

Explanations

phrases indicating a sequence of events with resulting consequences

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a sequence of events or actions

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

Problem

-0.66

floor

-0.65

BLE

-0.63

worn

-0.63

vent

-0.63

ve

-0.62

isp

-0.62

ut

-0.62

Sty

-0.60

 Unch

-0.60

POSITIVE LOGITS

soever

0.87

akespeare

0.71

psons

0.69

 Kira

0.65

upon

0.65

>[

0.65

xual

0.64

ãĥĢ

0.64

eway

0.62

 Aviv

0.62

Activations Density 0.031%

phrases indicating a sequence of events with resulting consequences

phrases that indicate a sequence of events or actions

No Comments

No Known Activations