INDEX

Explanations

phrases involving the abbreviation "WH"

oai_token-act-pair · gpt-3.5-turbo

the term "WH" and its variations, possibly indicating references to specific types of questions or 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

zzo

-0.80

cess

-0.73

taboola

-0.72

amination

-0.71

bidden

-0.70

atro

-0.70

uador

-0.69

beans

-0.69

adic

-0.67

otation

-0.66

POSITIVE LOGITS

ILE

0.94

LY

0.93

YY

0.92

ICAL

0.84

ALE

0.84

LAN

0.84

OA

0.83

AX

0.82

HL

0.82

FIX

0.82

Activations Density 0.030%

phrases involving the abbreviation "WH"

the term "WH" and its variations, possibly indicating references to specific types of questions or contexts

No Comments

No Known Activations