INDEX

Explanations

numerical terms related to seasons, volumes, weeks, and groups

oai_token-act-pair · gpt-3.5-turbo

references to seasons and episodes in series or shows

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uckland

-0.73

ught

-0.70

alty

-0.67

ndra

-0.66

issance

-0.61

asking

-0.61

ifles

-0.60

 Aram

-0.59

eport

-0.58

ousing

-0.58

POSITIVE LOGITS

VII

1.23

III

1.22

 VIII

1.22

XII

1.15

IV

1.14

1.11

1.10

1.08

1.07

Activations Density 0.339%

numerical terms related to seasons, volumes, weeks, and groups

references to seasons and episodes in series or shows

No Comments

No Known Activations