INDEX

Explanations

references to specific job titles or roles

oai_token-act-pair · gpt-3.5-turbo

empty or placeholder tokens

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

selves

-0.67

fw

-0.60

ĸļ

-0.60

ankind

-0.59

antics

-0.59

burning

-0.59

escal

-0.58

ickle

-0.56

ciating

-0.55

apo

-0.55

POSITIVE LOGITS

 extraord

0.72

 Jeremy

0.63

 defends

0.63

Joined

0.63

 councillor

0.61

 Shaun

0.61

 warns

0.61

sych

0.61

 Erik

0.60

 defended

0.58

Activations Density 0.315%

references to specific job titles or roles

empty or placeholder tokens

No Comments

No Known Activations