INDEX

Explanations

references to specific media channels or networks

oai_token-act-pair · gpt-3.5-turbo

references to television networks and shows

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

iasis

-1.04

olulu

-0.77

heid

-0.71

ogy

-0.70

arial

-0.69

antly

-0.69

staking

-0.69

emort

-0.67

raged

-0.67

ority

-0.65

POSITIVE LOGITS

trop

0.94

NT

0.89

KI

0.88

OULD

0.87

OW

0.87

ND

0.86

CW

0.86

soDeliveryDate

0.83

RC

0.82

VR

0.82

Activations Density 0.017%

references to specific media channels or networks

references to television networks and shows

No Comments

No Known Activations