INDEX

Explanations

negative events or outcomes described with strong adjectives

oai_token-act-pair · gpt-3.5-turbo

terms related to negative consequences or failures

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

 weights

-0.71

encers

-0.71

activated

-0.71

ingham

-0.71

llular

-0.71

ju

-0.67

agonal

-0.67

alde

-0.66

strength

-0.65

ants

-0.65

POSITIVE LOGITS

itous

0.84

 miser

0.81

 disastrous

0.77

 Spac

0.75

 mishand

0.73

ously

0.72

 Ples

0.71

ilton

0.70

 disappoint

0.67

erella

0.67

Activations Density 0.033%

negative events or outcomes described with strong adjectives

terms related to negative consequences or failures

No Comments

No Known Activations