INDEX

Explanations

mentions of a specific media outlet or publication

oai_token-act-pair · gpt-3.5-turbo

references to the word "Fam" or variations of it in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

dress

-0.74

 Wonderland

-0.71

 skirts

-0.71

graduate

-0.69

 overs

-0.68

 dwarf

-0.68

 tint

-0.68

pine

-0.66

İĭ

-0.66

DOWN

-0.65

POSITIVE LOGITS

ilial

1.69

iliar

1.53

ilitation

1.12

ilar

1.05

igl

1.04

uci

0.98

ilit

0.97

ilies

0.95

itsu

0.95

ili

0.95

Activations Density 0.010%

mentions of a specific media outlet or publication

references to the word "Fam" or variations of it in different contexts

No Comments

No Known Activations