INDEX

Explanations

adverbs that emphasize a negative evaluation or describing something as in a poor state

oai_token-act-pair · gpt-3.5-turbo

negative descriptions or critiques

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

raltar

-0.78

iferation

-0.72

inarily

-0.72

ership

-0.71

heny

-0.71

htaking

-0.69

iture

-0.68

anwhile

-0.68

itures

-0.67

itionally

-0.67

POSITIVE LOGITS

 behaved

0.93

 suited

0.83

 enough

0.81

asses

0.78

 mistaken

0.75

 damaged

0.73

 beaten

0.73

needed

0.70

 patched

0.69

 Stab

0.69

Activations Density 0.020%

adverbs that emphasize a negative evaluation or describing something as in a poor state

negative descriptions or critiques

No Comments

No Known Activations