INDEX

Explanations

mentions of specific geographical and political terms

oai_token-act-pair · gpt-3.5-turbo

references to locations and related events

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

igans

-0.79

urally

-0.70

hend

-0.70

UF

-0.69

anz

-0.68

gn

-0.67

fig

-0.67

istance

-0.67

eric

-0.65

DCS

-0.65

POSITIVE LOGITS

 Thumbnails

1.04

taboola

0.93

[|

0.79

Untitled

0.77

isSpecialOrderable

0.72

NetMessage

0.71

 exha

0.69

 antioxid

0.68

 Pradesh

0.67

 upkeep

0.66

Activations Density 0.075%

mentions of specific geographical and political terms

references to locations and related events

No Comments

No Known Activations