INDEX

Explanations

words related to negative actions or qualities, such as discrediting, distaste, or leniency

oai_token-act-pair · gpt-3.5-turbo

key terms related to judgment and evaluation

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

Kit

-0.87

 Kits

-0.80

BI

-0.77

 Â®

-0.77

meg

-0.75

 Helic

-0.74

Cu

-0.73

iP

-0.70

 TECH

-0.70

HS

-0.69

POSITIVE LOGITS

itionally

1.02

ested

1.01

animous

0.99

unction

0.97

irable

0.94

eful

0.94

empt

0.93

atically

0.93

iless

0.91

iencies

0.91

Activations Density 0.308%

words related to negative actions or qualities, such as discrediting, distaste, or leniency

key terms related to judgment and evaluation

No Comments

No Known Activations