INDEX

Explanations

instances where something is explicitly made clear or declared

oai_token-act-pair · gpt-3.5-turbo

the word "it" in various contexts, often indicating emphasis or a key subject matter

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Shares

-0.65

atur

-0.60

idth

-0.57

pta

-0.57

wcsstore

-0.55

 Balloon

-0.55

 treasure

-0.54

ruction

-0.54

 funer

-0.53

è¦

-0.53

POSITIVE LOGITS

 easier

0.97

 happen

0.93

 abundantly

0.93

 impossible

0.80

 harder

0.76

 seem

0.76

 clearer

0.74

unes

0.73

 clear

0.73

 safer

0.73

Activations Density 0.050%

instances where something is explicitly made clear or declared

the word "it" in various contexts, often indicating emphasis or a key subject matter

No Comments

No Known Activations