INDEX

Explanations

names with "hart" or "hardt"

oai_token-act-pair · gpt-3.5-turbo

references to specific leaders and their associated names

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

Magikarp

-0.70

URI

-0.69

 millenn

-0.68

ntil

-0.67

 veter

-0.65

 tampering

-0.65

 sclerosis

-0.65

pty

-0.63

 corros

-0.61

 sarc

-0.60

POSITIVE LOGITS

hart

0.97

awi

0.93

ford

0.85

ãĤ¤ãĥĪ

0.82

enstein

0.81

hardt

0.80

 Ducks

0.79

enberg

0.76

ilton

0.75

Jr

0.73

Activations Density 0.019%

names with "hart" or "hardt"

references to specific leaders and their associated names

No Comments

No Known Activations