INDEX

Explanations

the name "Blake" at varying levels of relevance

oai_token-act-pair · gpt-3.5-turbo

repetitions of the name "Blake" in various 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

egal

-0.78

ocard

-0.76

ktop

-0.75

Demand

-0.74

CHO

-0.72

ãĥķãĤ©

-0.70

oppable

-0.69

ocre

-0.68

OLOG

-0.68

nih

-0.67

POSITIVE LOGITS

pedia

0.84

 Blake

0.83

rette

0.81

leans

0.78

osure

0.74

querque

0.74

 Griffin

0.73

more

0.72

antine

0.70

Isa

0.69

Activations Density 0.009%

the name "Blake" at varying levels of relevance

repetitions of the name "Blake" in various contexts

No Comments

No Known Activations